This is a first time attempt at object detection in C++. The main file is where the object detection is written. Within main it loads a YOLOv5 nano model using
cv::dnn::readNet(...). Then runs inference on either a single image or on video feed, the user is given the option to pick. Once the inference is run, it maps the bounding boxes on the feed/input and then displays it.
I wrote this software to improve my general C++ skills and learn more about object detection inside this language. I learned about NMS, object detection workflow, some new OpenCV functions, and how to work with video feed.
This project used Visual Studio and OpenCV to complete.
These items need to be fixed in future updates.
- Video feed needs to update overtime.
- Implement this into command line arguments.
- Add more comments to code explaining each step.